Learning to Detect Mobile Objects from LiDAR Scans Without Labels

TL;DR

This paper introduces a label-free method for training 3D object detectors from LiDAR scans by using heuristics to generate initial seed labels and self-training to achieve high accuracy without human annotations.

Contribution

It presents a novel approach that leverages simple heuristics and self-training to detect mobile objects in LiDAR data without any labeled training data.

Findings

Achieves high detection accuracy using unlabeled data.

Effectively bootstraps detectors with heuristic-generated seed labels.

Eliminates the need for costly human annotations.

Abstract

Current 3D object detectors for autonomous driving are almost entirely trained on human-annotated data. Although of high quality, the generation of such data is laborious and costly, restricting them to a few specific locations and object types. This paper proposes an alternative approach entirely based on unlabeled data, which can be collected cheaply and in abundance almost everywhere on earth. Our approach leverages several simple common sense heuristics to create an initial set of approximate seed labels. For example, relevant traffic participants are generally not persistent across multiple traversals of the same route, do not fly, and are never under ground. We demonstrate that these seed labels are highly effective to bootstrap a surprisingly accurate detector through repeated self-training without a single human annotated label.